Functional Characterization of High Levels of Meningioma 1 As Collaborating Oncogene in Acute Leukemia

Functional characterization of high levels of meningioma 1 as collaborating oncogene in acute leukemia

T Liu1, D Jankovic1, L Brault1, S Ehret1, F Baty1, V Stavropoulou1, V Rossi2, A Biondi2 and J Schwaller1

1Department of Biomedicine, University Hospital Basel, Basel, Switzerland and 2The Centro M. Tettamanti, Clinica Pediatrica, Universita` Milano-Bicocca, Monza, Italy

Retroviral expression of leukemogenic oncogenes in the system by a conditional knock-in strategy resulted in the murine hematopoietic system is essential but not sufficient to formation of T-cell lymphomas as well as AML after a long induce acute leukemia. Proviral integration-mediated elevated latency, suggesting that MN1-TEL, similar to MLL-X fusions, is expression of the meningioma 1 (MN1) oncogene suggested 4–6 MN1 acting as cooperating event in mixed-lineage leukemia 1 essential but not sufficient to induce the disease. Gene (MLL) and eleven nineteen leukemia (ENL)-induced murine expression profiling studies of a large number of human leukemia leukemia. Indeed, co-expression of MN1 with MLL-ENL en- samples showed that MN1 is deregulated in cases with alterations hanced transformation in vivo, and resulted in a significantly at 3q26 leading to ecotropic virus integration-1 (EVI1) over- reduced latency for induction of an aggressive acute leukemia expression.7 In addition, elevated MN1 expression has been when compared with MN1 or MLL-ENL alone. In addition, co- associated with the presence of inv16 leading to a core-binding expression of MN1 increased the granulocyte macrophage 8 progenitor cell population with leukemia-initiating properties as factor-b/MYH11 fusion. Recently, functional studies have shown shown in secondary transplantation experiments. Gene expres- that overexpression of MN1 alone is able to induce an AML sion profiling experiments identified putative downstream phenotype in mice.8,9 Furthermore, high MN1 expression was MN1 targets, of which FMS-like tyrosine kinase 3 (FLT3) and shown to have negative prognostic effect in AML, especially in CD34 were upregulated in both MN1-overexpressing murine the absence of common karyotype abnormalities.10,11 leukemias and in pediatric acute leukemias with high MN1 levels. Interestingly, small interfering RNA (siRNA)-mediated Recurrent chromosomal translocations resulting in expression MN1 knockdown resulted in cell cycle arrest and impaired clono- of oncogenic fusion genes are the hallmark of human genic growth of human leukemia cell lines with high MN1 levels. leukemias.12 Mixed-lineage leukemia 1 (MLL1) on 11q23 Our work shows for the first time that high MN1 levels are is one of the most frequently altered genes in human leukemia important for the growth of leukemic cells, and that increased with MLL fusion genes involving over 50 partners or partial MN1 expression can synergize with MLL-ENL and probably tandem duplications.13,14 Several studies have shown that other transforming fusion genes in leukemia induction through a distinct gene expression program that is able to expand the a large number of MLL fusions as well as MLL-partial leukemia-initiating cell population. tandem duplications have in vitro and/or in vivo transforming Leukemia (2010) 24, 601–612; doi:10.1038/leu.2009.272; activity. Indeed, transplantation of bone marrow cells retro- published online 14 January 2010 virally expressing various MLL fusions (such as MLL-AF9, Keywords: meningioma1; acute leukemia; oncogene; collaboration MLL-ENL or MLL-GAS7) led to the induction of a lethal disease mimicking human acute leukemia. The clonal character as well as a latency period of 60 to 4200 days of the induced leukemia suggested that although MLL-X fusions are essential, they might Introduction not be fully sufficient for inducing a leukemic phenotype in vivo.15–19 The meningioma 1 (MN1) gene was first identified as the target Elevated MN1 expression was not only associated with the of a sporadic balanced chromosomal translocation in a patient presence of distinct genetic alterations (e.g. inv16, alterations of with meningioma.1 The absence of MN1 expression in the index the EVI1 gene locus) in human leukemia, but the MN1 gene patients has led to the suggestion that MN1 is a candidate tumor locus was also target of proviral integration site in mouse suppressor gene. Several studies have proposed that MN1 leukemia models induced by bone marrow reconstitution of retrovirally expressed oncogenes, such as mutant AML1 or presumably exerts an effect as a transcriptional cofactor, most 20,21 probably through interaction with other transcriptional regula- NUP98-HOXD13. Searching for potentially cooperating tors, such as p300/CREB-binding protein.2 MN1 was first linked events by cloning proviral integration sites in murine leukemia to human leukemia after the cloning of the balanced chromo- induced by retrovirally expressed MLL-ENL, we found increased somal translocation t(12;22)(p13;q12) in patients with acute MN1 expression in leukemic cells harboring an integration near myeloid leukemia (AML), myelodysplasia or chronic myelo- the MN1 locus. These observations suggested that elevated genous leukemia. This translocation leads to the expression of MN1 levels might synergize in leukemia induction initiated by an MN1–TEL fusion that consists of almost the entire open distinct genetic alterations such as MLL-ENL. By co-expression reading frame of MN1 fused with the DNA-binding moiety of of MN1 and MLL-ENL in the murine hematopoietic system, we TEL.3 Expression of MN1–TEL in the mouse hematopoietic were able to show that overexpression of MN1 can exert an effect as a collaborative genetic hit in MLL-ENL-induced leukemia in vivo through expansion of the pool of leukemia- Correspondence: Professor J Schwaller, Department of Biomedicine, initiating cells. New putative MN1 target genes were identified University Hospital Basel, Hebelstrasse 20, ZLF, Lab 318, Basel CH-4031, Switzerland. E-mail: [email protected] and validated in mouse and human leukemias overexpressing Received 25 May 2009; revised 2 November 2009; accepted 23 MN1. For the first time, we show that small interfering RNA November 2009; published online 14 January 2010 (siRNA)-mediated knockdown of MN1 impaired proliferation of Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 602 human leukemic cells with abundant levels of MN1, suggesting days. Colonies were harvested, and 104 cells were replated in that MN1 could represent a new therapeutic target. methylcellulose for four rounds.

Material and methods Analysis of transplanted mice After red cell lysis peripheral blood and bone marrow cells were Cell lines counted and analyzed using a flow cytometer (Cyan II, Becton The following human leukemia cell lines were analyzed: Dickinson, Franklin Lakes, NJ, USA), single-cell suspensions MV4;11, acute myeloid leukemia (MLL-AF4 þ , FMS-like were stained with phycoerythrin, or allophycocyanin fluoro- tyrosine kinase 3 (FLT3)-ITD þ ), MOLM13, acute myeloid chromes-labeled c-Kit, Sca1, Gr1, Mac1, B220 and CD34 leukemia (MLL-AF9 þ ); EOL1, acute eosinophilic leukemia monoclonal antibodies (all from Pharmingen, San Diego, CA, (MLL-partial tandem duplication); THP1, acute monocytic USA). Histopathological analysis of peripheral blood, bone leukemia (MLL-AF9 þ ); KOCL44, acute lymphoblastic leukemia marrow and hematopoietic organs were performed using (MLL-ENL þ ), SEM, acute lymphoblastic leukemia (MLL- standard procedures. AF4 þ ), KOPN8, acute lymphoblastic leukemia (MLL-ENL þ ), RS4;11, acute lymphoblastic leukemia (MLL-AF4 þ ) and HL-60, acute myeloid leukemia. All cells were kept in RPMI-1640 with Southern blot analysis Southern blot analysis was performed using standard protocols.22 Glutamine (Invitrogen, Carlsbad, CA, USA) plus 10% fetal Clonality of the MSCV-MLL-ENL provirus was assessed using a bovine serum and penicillin/streptomycin at 37 1C. 2.2 kb Hind III human MLL1 cDNA fragment as the hybridization probe.23 The Mn1 locus-specific probe was generated using PCR Construction of recombinant retroviral vectors and the primer sequences are available upon request. Full-length human 50-FLAG-tagged MN1 complementary DNA (cDNA) was excised (SacI-HindIII) from pCMVTag2B-MN1 (a kind Retroviral integration cloning by splinkerette PCR gift from Dr Paul MacDonald, Cleveland, OH) and transferred into Genomic DNA isolated from bone marrow or spleen cells of the pSL1180 (SacI-SmaI) cloning vector and further subcloned into leukemic mice was digested with NlaIII or MseI for 12–16 h, and pMSCV-IRES/EYFP.TheMLL-ENLcDNA(akindgiftfromDrRobert ligated to the splinkerette linker overnight. The nested PCR was Slany, Erlangen, Germany) was subcloned from pMSCV-pgk/neo performed by using splinkerette linker-specific primers and into pMSCV-IRES/EGFP (for single transductions) or pMSCV- primers recognizing the long terminal repeat of pMSCV.24,25 pgk/puro (for double transductions) using a unique EcoRI site. All Amplicons from the second PCR were separated on 2% agarose expression plasmids were verified by extensive restriction digests, gel, purified by gel purification kit (Qiagen, Hilden, Germany) and and by partial sequencing. subcloned into pCR 2.1-TOPO vector (Invitrogen) before sequencing, or directly sequenced by using the BigDye Terminator v3.1 chemistry and ABI 3130 DNA genetic analyzer (Applied Bone marrow infections and transplantation Biosystems, Foster City, CA, USA). The obtained sequences were Bone marrow cells were harvested from 6- to 10-week-old (FVB/ analyzed using BLAST against the mouse genome database of the N Â 129/S1)F1 mice 4 days after intraperitoneal injection of National Center for Biotechnology Information. 5-fluorouracil 150 mg/kg (Sigma, St Louis, MO, USA), and were cultured for 24 h in RPMI-1640 supplemented with 10% fetal bovine serum, 10 ng/ml of human interleukin-6 (IL-6), 6 ng/ml of siRNA knockdown experiments murine IL-3 and 100 ng/ml of murine stem cell factor The MN1-specific short hairpin RNA or scramble short hairpin (PeproTech EC, London, UK). HEK293T cells were transiently RNA lentiviral plasmids (pLKO.1-puro) were purchased from co-transfected with pMSCV retroviral vector and a packaging Sigma (cat. no. 07310707MN). Lentiviral vector packaging was vector (pIK6), and virus containing supernatants were collected performed according to the calciumphosphate method. In brief, after 48 h and concentrated by 60 min of centrifugation at short hairpin RNA lentiviral plasmid was mixed with envelope 14 000 r.p.m. at 4 1C. Retroviral infections were performed by plasmid pMD2G, packaging plasmid pMDLg/pRRE and Rev- 1 spinoculation of the cells (2500 r.p.m. for 90 min at 30 C) with expression plasmid pRSV-Rev, and 0.5 M CaCl2 was added. The 6 retroviral supernatant on two consecutive days. 1 Â 10 mixture was added drop-wise to 2 Â BES (N,N-bis(2-hydroxy- transduced bone marrow cells were injected into the tail vein ethyl)-2-aminoethanesulfonic acid) while vortexing at full 60 of lethally irradiated ( Co, 950 rad) syngenic recipients. speed. After 20 min of incubation, the precipitate was slowly Transduction with two viruses (MLL-ENL in pMSCV-puro and added to the 293T cell monolayer. Lentiviral supernatants were MN1 in pMSCV-YFP) was performed sequentially: after trans- harvested after 48 h by spinning at 2500 r.p.m. and filtering duction with the MLL-ENL-expressing virus, the cells were through 0.45 mm filter. Lentivirus were further concentrated by selected for 7 days before transduction with the MN1-encoding spinning through Vivaspin 20 concentrator (Sartorius Biolabs, virus. For secondary transplantation, the limiting doses of Goettingen, Germany) at 4000 r.p.m. at 4 1C for 30 min, and 5 leukemic blasts (from 5 Â 10 to 500) from primary mice were stored at À80 1C. transplanted into sublethally irradiated (450 rad) recipients by tail vein injection. Immunofluorescence Cells were first fixed with 4% paraformaldehyde and then Serial replating assay cytospin and immunolocalization were performed. Cells were 104 transduced bone marrow cells were plated in 1 ml of permeabilized with 0.5% Triton X-100 in phosphate-buffered methylcellulose culture (Stem Alpha.mIE, Stem Alpha, St saline, treated with RNase and blocking was performed in 0.1% Clement les Places, France) containing IL-3, IL-6 and murine Tween-20/phosphate-buffered saline supplemented with 1% stem cell factor. The number of colonies was counted after 7 bovine serum albumin. The MN1 primary antibody (a kind gift

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 603 from M Meesters and E Zwarthoff, Rotterdam, The Netherlands) were processed with an Affymetrix GeneChip Scanner 3000 7G. was added for 1 h incubation at room temperature, and slides DAT image files of the microarrays were generated using were then washed and incubated in the dark for 1 h with the Affymetrix GeneChip Command Console (AGCC, version Alexa Fluor 488 goat anti-mouse antibody (Invitrogen). Propi- 0.0.0.676). All statistical analyses were performed using Gene- dium iodide was used for nuclei staining. Slides were washed Spring GX software (Agilent). Genes were considered as and mounted with Fluorsafe Reagent (Calbiochem, San Jose, significant whenever the fold change was superior to 1.5 and CA, USA). Confocal microscopy was carried on a LSM 510 the P-value of o0.05. The data discussed in this publication laser-scanning microscope (Zeiss, Oberkochen, Germany). have been deposited in the Gene Expression Omnibus (http:// www.ncbi.nlm.nih.gov/geo/) of the National Center for Biotechnology Information and will be accessible through Fluorescence-activated cell sorting of GMPs accession number GSE13189. The isolated bone marrow cells were first stained with the lineage cocktail that contains antibodies specific for the following mouse lineage markers: biotinylated rat anti-mouse Patient samples CD5, CD11b, CD45R/B220, Ly-6G (Gr-1) and Ter119 Expression of MN1 and putative downstream targets were (MAGM209, R&D Systems, Minneapolis, MN, USA). Cells were measured in a panel of childhood acute leukemia bone marrow then stained with a streptavidin Pacific Blue-conjugated samples. Under informed consent by the guardians, 21 children (Invitrogen), a phycoerythrin-Cy5-conjugated anti-mouse diagnosed with acute lymphoblastic leukemia (ALL) and 6 with IL-7Ra (A7R34, eBioscience, San Diego, CA , USA), a phyco- AML, according to the conventional FAB (French American erythrin-conjugated anti-mouse FcgRII/III (93, eBioscience), an British) and immunological criteria, were included. In all, 8 Alexa Fluor 647-conjugated anti-mouse CD34 (RAM34, patients who were diagnosed of ALL were o1 year of age. On BD Biosciences, San Jose, CA, USA), an allophycocyanin- the basis of the immunophenotype, 15 patients were classified conjugated anti-mouse c-Kit (2B8, BD Pharmigen) and a as B-cell precursor ALL, and 6 as T-ALL. Chromosome phycoerythrin-Cy7-conjugated anti-mouse Sca-1 (E13-161.7, suspensions were used for cytogenetic and fluorescence in situ BioLegend, San Diego, CA, USA) monoclonal antibody. hybridization analyses. All infant leukemia samples harbored Granulocyte macrophage progenitors (GMPs) were fluores- 11q23 alterations. Among the B-cell ALL patients were two cence-activated cell sorted as IL-7RaÀ LinÀ Sca-1À c-Kit þ patients with t(1;19) and one patient with t(12;21). One AML CD34 þ FcgRII/IIIhigh as described previously.16 patient was positive for t(8;21). For all samples, mononuclear cells were routinely isolated using Ficoll-Hypaque (Pharmacia- LKB, Uppsala, Sweden) density gradient centrifugation (bone Quantitative reverse transcriptase-PCR marrow mononuclear cells) and then RNA isolated. All samples Target validation was performed in triplicate by real-time PCR contained 490% of blasts as assessed by morphological and with SYBR-green on an ABI prism 7700 sequence detection immunological criteria (not shown). system (Applied Biosystems). For each target the results were normalized to glyceraldehyde-3-phosphate dehydrogenase and given as DDCt values normalized to vector (mock)-infected Results (MSCV-IRES/EYFP) bone marrow cells. The results are repre- sented as mean±s.d. of three independent experiments. The While searching for potentially cooperating events by studying detailed information about oligonucleotide primers and condi- viral integration sites in murine leukemia induced by reconstitu- tions used are available upon request. tion with bone marrow cells that retrovirally expressed the MLL- ENL fusion, we found MN1 as an integration site flanking gene. As shown in Table 1, we cloned 25 integrations from 10 mouse Gene expression profiling MLL-ENL leukemias. From leukemic blasts of three animals, only In three independent experiments, bone marrow cells were a single integration was found, whereas 5 others harbored 3 or 4 transduced with MSCV-MN1-IRES/EYFP or empty vector. At integrations. Southern blot analysis further confirmed the (oligo)- 72 h after transduction, EYFP-positive cells were sorted with clonal nature of the disease and the provirus insertion adjacent to fluorescence-activated cell sorting and RNA was isolated using the Mn1 gene locus in one sample (‘ME10’) (Supplementary ion-exchange chromatography with RNAmini (Qiagen) accord- Figure 1). Integration near the MN1 locus resulted in significant ing to the manufacturer’s protocol. cDNA target was synthe- upregulation of its mRNA expression in the respective animal sized, fragmented, biotin-labeled using the Whole Transcript (Figure 1a). We therefore decided to study the effect of elevated Target Labeling and Control Reagents (Affymetrix, cat. no. MN1 expression in MLL-ENL-mediated leukemia. 900652, Santa Clara, CA, USA) starting from 200 ng total RNA, We first studied the transforming potential of elevated MN1 according to the procedure described in the Affymetrix expression in primary mouse bone marrow cells. Retroviral GeneChip Whole Transcript Sense Target Labeling Assay expression of human MN1 resulted in blocked differentiation, Manual, Version 4. The cDNA was fragmented and the resulting aberrant self-renewal and growth advantage, as shown in serial fragments of approximately 40–70 nucleotides were monitored replating assays in methylcellulose and in liquid cultures with the Bioanalyzer using the RNA Nano 6000 Chip (Agilent, confirming previous observations8,9 (data not shown). This Palo Alto, USA). The hybridization cocktail containing suggested that MN1 might functionally collaborate with the fragmented biotin-labeled target DNA at a final concentration MLL-ENL fusion in the development of the acute leukemia of 25 ng/ml was transferred into GeneChip Mouse Gene 1.0 ST phenotype. To experimentally address this hypothesis, we Arrays (Affymetrix, cat.no. 901168) and incubated at 45 1Cona performed a series of bone marrow transplant experiments, in rotator in a hybridization oven 640 (Affymetrix) for 17 h at which we reconstituted lethally irradiated mice with bone 60 r.p.m. The arrays were washed and stained on a Fluidics marrow transduced with retrovirus expressing MLL-ENL, MN1 Station 450 (Affymetrix) using the Hybridisation Wash and Stain or both. For co-expression bone marrow cells were first Kit using the Fluidics Procedure FS450_0007. The GeneChips transduced with the virus encoding MLL-ENL, followed shortly

Leukemia 604 Leukemia

Table 1 Retroviral integration sites cloned from 10 murine leukemias induced by MLL-ENL

Mouse Gene Full name (Predicted) gene function Location and distance Orientation Mouse chr. Human chr. CISa no. symbol for retroviral integration

ME 1 Ib Rps6kb1 Ribosomal protein S6 kinase polypeptide 1 Protein serine/threonine kinase Downstream 5659 bp Opposite 4 17q23.1 NA Hmgn2 High mobility group nucleosomal binding domain 2 DNA binding protein Upstream 76 653 bp Opposite 4 1p36.1 NA II Supt4h2 Suppressor of Ty 4 homolog 2 Transcription factor Downstream 6797 bp Opposite 11 17q21-q23 NA Bzrap1 Benzodiazapine receptor associated protein 1 Benzodiazepine receptor binding Upstream 29 292 bp Opposite 11 17q22-q23 NA

III Clec16a C-type lectin domain family 16, member A Sugar binding Downstream 28 879 bp Same 16 16p13.13 1 oncogene leukemogenic collaborating as 1 Meningioma Socs1 Suppressor of cytokine signaling 1 Negetive regulator of JAK-STAT signaling Upstream 13 327 bp Opposite 16 16p13.13 NA ME 2 Saps1 SAPS domain family, member 1 Regulation of phosphoprotein Downstream 5254 bp Same 7 19q13.42 2 phosphatase activity Hspbp1 SPA) binding protein, cytoplasmic cochaperone 1 hsp70-interacting protein Upstream 8554 bp Same 7 19q13.42 NA ME 3 I Flt3l FMS-like tyrosine kinase 3 ligand Ligand of receptor protein tyrosine kinase Downstream 541 bp Opposite 7 19q13.3 NA Aldh16a1 Aldehyde dehydrogenase 16 family, member A1 Oxidoreductase activity Upstream 5308 bp Opposite 7 19q13.33 3 II Cd83 CD83 antigen Adhesion receptor Downstream 633 289 bp Same 13 6p23 1 Jarid2 Jumonji protein DNA binding protein Upstream 293 541 bp Same 13 6p24-p23 NA III Clrn3 Clarin 3 Transmembrane protein Downstream 55 148 bp Opposite 7 10q26.2 4 Ptpre Protein tyrosine phosphatase, receptor type, E Protein tyrosine phosphatase Upstream 60 092 bp Same 7 10q26 4 IV Ralgps2 Ral GEF with PH domain and SH3 binding motif 2 Guanyl-nucleotide exchange factor Downstream 73 087 bp Opposite 1 1q25.2 NA Angptl1 Angiopoietin-like 1 Receptor protein tyrosine kinase Downstream 126 043 bp Same 1 1q25.2 NA signaling pathway 1700057K13Rik Hypothetical protein LOC73435 Unknown Upstream 112 647 bp Opposite 1 1q25.2 NA ME 4 I Hsp90aa1 Heat shock protein 1, alpha Molecular chaperone Downstream 4932 bp Opposite 12 14q32.33 NA Wdr20a WD repeat domain 20 Unknown Upstream 37 074 bp Same 12 N/A NA II Mid1ip1 Mid1 interacting protein Negative regulator of microtubule Downstream 940 212 bp Opposite X Xp11.4 NA depolymerization Liu T Bcor BCL-6 interacting corepressor isoform a Transcription repressor Upstream 371 669 bp Same X Xp21.2–p11.4 9

III Tnfrsf1b Tumor necrosis factor receptor superfamily, member 1b Tumor necrosis factor receptor 1st intron Same 4 1p36.3-p36.2 1 al et IV Itm2b Integral membrane protein 2B Induction of apoptosis Downstream 10 455 bp Opposite 14 13q14.3 NA Med4 Vitamin D receptor interacting protein Transcription regulator Upstream 113 900 bp Same 14 13q14.2 NA ME 5 I Ly86 Lymphocyte antigen 86 Immune response Downstream 204 132 bp Same 13 6p25.1 2 Rreb1 Ras responsive element binding protein 1 isoform 2 Transcription regulator Upstream 271 015 bp Same 13 6p25 9 II Psd2 Pleckstrin and Sec7 domain containing 2 ARF guanyl-nucleotide Downstream 181 588 bp Opposite 18 5q31 1 exchange factor Pura Purine rich element binding protein A DNA binding protein Upstream 92 667 bp Opposite 18 5q31.3 1 ME 6 I Rbks Ribokinase Ribose kinase 7th intron Same 5 2p23.3 NA II Retn Resistin Hormone Downstream 5657 bp Same 8 19p13.2 NA 1810033B17Rik Hypothetical protein LOC69189 Unknown Upstream 2515 bp Same 8 19p13.2 NA III Noxa1 NADPH oxidase activator 1 Superoxide-generating NADPH Downstream 69 311 bp Same 2 9q34.3 NA oxidase activator Nrarp Notch-regulated ankyrin repeat protein Notch signaling pathway Upstream 16 770 bp Opposite 2 9q34.3 NA ME 7 Malat1 Metastasis associated lung adenocarcinoma Prognostic parameter for nonsmall cell Downstream 5581 bp Same 19 11q13.1 NA transcript 1 (non-coding RNA) lung cancer Frmd8 FERM domain containing 8 Cytoskeleton Upstream 49 265 bp Same 19 11q13 3 ME 8 Gadd45b Growth arrest and DNA-damage-inducible 45 beta MAPK signaling pathway Downstream 9864 bp Opposite 10 19p13.3 1 Gng7 Guanine nucleotide binding protein (G protein), G-protein coupled receptor signaling pathway Upstream 9958 bp Same 10 19p13.3 NA gamma 7 subunit ME 9 I Slc38a2 Solute carrier family 38, member 2 Amino acid and ion transport Downstream 61 937 bp Opposite 15 12q 6 Slc38a4 Solute carrier family 38, member 4 Amino acid and ion transport Upstream 236 288 bp Opposite 15 12q13 NA II Tcf7 Transcription factor 7, T-cell specific WNT signaling pathway 1st intron Same 11 5q31.1 NA ME 10 I Mgat5 Mannoside acetylglucosaminyltransferase 5 Acetylglucosaminyltransferase 1st intron Same 1 2q21 1 II LOC629605 Hypothetical protein LOC629605 Unknown Downstream 8254 bp Same 12 N/A NA Ppp2r5c Protein phosphatase 2, regulatory subunit B (B56), Protein phosphatase type 2A regulator Upstream 26 735 bp Same 12 14q32 3 gamma isoform III Socs7 Suppressor of cytokine signaling 7 Negative regulation of insulin receptor Downstream 12 212 bp Opposite 11 17q12 NA signaling pathway LOC665512 Similar to Rho GTPase activating protein 21 Unknown Upstream 38 114 bp Opposite 11 N/A NA IV Mn1 Meningioma 1 Oncogene Downstream 99 834 bp Same 5 22q11 8 C130026L21Rik Hypothetical protein LOC330164 Unknown Upstream 27 121 bp Same 5 N/A NA Abbreviations: CIS, common integration sites; chr., chromosome; ME, murine leukemia induced by overexpression of MLL/ENL; NA, not applicable. aThe numbers in CIS indicate hits of the same integration flanking gene found in RTCGD database (http://rtcgd.abcc.ncifcrf.gov/). bRoman numerals in ME mean different insertion sites identified from the same leukemic mouse. Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 605

Figure 1 Identification of MN1 as putative cooperative oncogene in MLL-ENL-induced mouse leukemia. (a) mRNA expression of MN1 in murine leukemia induced by retroviral expression of the MLL-ENL fusion gene. Relative expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression in the same sample and standardized to gene expression in normal murine bone marrow (BM). The animal with the integration near the MN1 locus (black bar, ME10) showed significant elevated levels of MN1 mRNA expression. Results are mean±s.d. of triplicates. (b) Co-expression of MN1 and MLL-ENL-induced acute leukemia in mice after a significantly shorter latency (median latency, 35 days) than MLL-ENL (64 days) or MN1 alone (100 days). Significance of the differences in the latency periods between the corresponding groups was assessed using log-rank test (Po0.001). Co-expression of MLL-ENL with the empty vector (pMSCV-EGFP) did not change the latency. (c) Expression of cellular surface markers in leukemic blasts of murine acute leukemia induced by MLL-ENL, MN1 and MLL-ENL þ MN1 was analyzed using flow cytometry. MLL-ENL þ MN1-expressing leukemic blasts are characterized by high levels of Gr1/Mac1, c-kit and CD34. (d) Histological analysis showed the appearance of leukemic blasts in blood smears and extensive infiltration of leukemic blasts in multiple organs including liver, spleen, and skeletal muscle of mice co-transduced with MLL-ENL and MN1.

by a second transduction with the MN1-expressing virus. (Figure 1b). In contrast to MLL-ENL or MN1 alone, reconstitution Analogous to previous studies, reconstitution with MLL-ENL of bone marrow co-transduced with virus encoding MLL-ENL (n ¼ 10) or with MN1 (n ¼ 8) expressing bone marrow in and MN1 induced an aggressive AML-like disease after a recipient mice led to the development of an acute leukemia signiﬁcantly shorter median latency of 35 days after transplanta- phenotype after a median latency of 64 or 100 days, respectively tion (Po0.001, log-rank test) (Figure 1b). Leukemic blasts from

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 606 MLL-ENL þ MN1 disease expressed higher levels of c-Kit and Figure 2). No change in the latency of the disease development CD34, but decreased levels of Gr1/Mac1 when compared with was observed when bone marrow was co-transduced with MLL- blasts from MLL-ENL-induced disease (Figure 1c). The disease ENL virus and an empty control virus (MSCV-EGFP ) (Figure 1b). was characterized by very high white blood cell counts, the These in vivo observations suggested that MN1 could exert presence of leukemic blasts in peripheral blood and excessive an effect as functional collaborator to MLL-ENL in murine infiltration in multiple organs such as bone marrow, spleen and leukemogenesis. liver (Figure 1d and Supplementary Table 1). Expression of both Leukemic blasts overexpressing MLL-ENL and MN1 showed a oncogenes in leukemic blasts was validated by reverse significantly increased clonogenic growth potential in vitro transcriptase PCR in single colonies of MLL-ENL þ MN1 when compared with leukemic cells induced by MLL-ENL or leukemic blasts growing in methylcellulose (Supplementary MN1 alone (Figure 2a). To address the in vivo significance of

Figure 2 Co-expression of MN1 with MLL-ENL increases the leukemia-initiating cell pool. (a) Increased in vitro colony formation in methylcellulose cultures of blasts from leukemic mice induced by the co-expression of MN1 with MLL-ENL when compared with MLL-ENL or MN1 alone. The significance was assessed in two ways: analysis of variance (ANOVA) tests with Bonferroni correction. Results represent mean values±s.d. of triplicates for three different leukemic mice. (b) Kaplan–Meier survival plot of primary and secondary transplantation with limiting dose of blasts from MLL-ENL or MLL-ENL þ MN1 leukemia mice revealed the synergistic potential of MN1 in disease induction. Significance of the differences in latency periods of MLL-ENL and MLL-ENL þ MN1 leukemic mice transplanted with the same cell number was assessed for each dose of cells by log-rank test (Po0.01). (c) The leukemic GMP (L-GMP) population in MLL-ENL, MN1 and MLL-ENL þ MN1 leukemic mice were defined using flow cytometric analysis as IL7RÀ, LinÀ, c-Kit þ , Sca1À,FcgRII/IIIhigh and CD34 þ fraction. The frequency of L-GMP in MLL- ENL þ MN1 mice is approximately 8.4% of total bone marrow, whereas it is approximately 20 times lower (0.34%) in MLL-ENL mice. (d) Kaplan– Meier survival plot of recipient mice transplanted with GMP population isolated from MLL-ENL or MLL-ENL þ MN1 leukemia mice. Median latency for disease development after transplant of 1000 MLL-ENL GMPs was 47 days, and was 35 days for 1000 MLL-ENL þ MN1 GMPs. Significance of the differences in latency period between the two groups of mice was assessed using log-rank test (Po0.01).

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 607 this observation, we performed a series of secondary transplan- the expression levels of these genes in blasts from mouse tation experiments using flow-sorted EGFP þ or EYFP þ cells. leukemias induced by MN1, MLL-ENL or co-expression of MLL- As shown in Figure 2b, transplantation of 5 Â 105 MLL-ENL ENL and MN1. In most murine samples, elevated expression leukemic blasts into sublethally irradiated recipients resulted in levels of DLK1, CD34, FLT3 and HLF was significantly AML after a median latency of 37 days. In contrast, transplanta- associated with the presence of elevated expression of MN1 tion of the same number of MLL-ENL þ MN1 leukemic cells (Figure 3a). resulted in AML after a significantly reduced latency (20 days, To validate these putative MN1 target genes in human Po0.01, log-rank test). Further transplantations of dilutions of leukemia, we compared their expression in a set of pediatric leukemic blasts showed that in contrast to MLL-ENL þ MN1 acute leukemia cases (including 7 B-ALL, 6 AML, 8 infant leukemia, 2000 MLL-ENL blasts were not sufficient to induce the leukemias and 6 T-ALL) with elevated expression of MN1 versus disease in all the recipients. These experiments provided strong those with low or absent MN1 expression (as found in normal evidence that MN1 could functionally collaborate with MLL- bone marrow). High MN1 levels were found in B-cell ALL and ENL and suggested that MN1 overexpression might lead to cases of infant leukemias (Figure 3b). We found significantly expansion of a leukemia-initiating cell population. higher expression levels of CD34 (P ¼ 0.002) and FLT3 Previous studied suggested that MLL fusion genes not (P ¼ 0.03) in pediatric leukemia cases with significantly elevated only transform hematopoietic stem cells, but are also able MN1 levels than in cases with low or undetectable MN1 to block differentiation and to provide aberrant self-renewal to levels. Expression of DLK1 and HLF1 correlated with MN1 in committed hematopoietic progenitor cells, including GMPs, some cases, but was not significant as a cohort (data not shown). common myeloid progenitors CMP and/or common lymphoid In contrast to the murine samples (Figure 3a), no association progenitors.15–17,26 The significantly shorter latency in of MEIS1 expression with elevated MN1 levels could be conjunction with a more immature, ‘stem-cell like’ phenotype observed. leukemia suggested that overexpression of MN1 might lead Although elevated MN1 expression has been reported to to an expansion of the leukemic progenitor population hit by provide negative prognostic information in adult AML the MLL fusion gene. It is worth noting that in normal cases,10,11 the contribution of MN1 to the biology of any bone marrow the highest levels of MN1 expression were found leukemic cell clone remained unclear. To functionally study the in GMPs, which are amplified in several MLL-mediated role of MN1 in a human leukemia, we first determined MN1 transformations and might include the leukemia-initiating expression levels in a panel of acute leukemia cell lines (HL60, cells.8,15,16 Therefore, we first compared the GMP populations MV4;11, MOLM13, KOCL44, RS4;11, SEM, KOPN8, EOL-1 and in leukemic mice induced by MLL-ENL, MN1 or MLL- THP1). As shown in Figure 4a, MN1 expression was detected in ENL þ MN1. As shown in Figure 2c, leukemias induced by some cell lines, such as THP1, MOLM13, KOCL44 and RS4;11, MN1 and MLL-ENL þ MN1 were characterized by a significant but not in all cell lines harboring MLL fusion genes. Findings on expansion of the GMP population, and co-expression of MN1 the mRNA level were also reflected on the protein levels as with MLL-ENL increased the frequency of GMPs as 20 times shown by immunofluorescent staining of MN1 in MOLM13 more than MLL-ENL. versus SEM cells (Figure 4b). To study the role of elevated MN1 To test the possibility of an increase in leukemia-initiating expression for maintenance of the malignant cell phenotype, we GMPs in leukemic mice expressing MLL-ENL þ MN1, we transduced the cells with a lentivirus expressing either MN1- transplanted identical numbers of flow-sorted GMPs into specific or scramble siRNAs. A series of control experiments sublethally irradiated hosts. As shown in Figure 2d, transfer of showed that the four cell lines of the panel (HL60, MOLM13, 1000 GMPs from MN1 þ MLL-ENL leukemia reproduced the RS4;11 and THP1) could be efficiently transduced (not shown). leukemic phenotype in all recipients after a median latency of Expression of MN1 siRNAs resulted in a reduction of MN1 35 days. In contrast, an identical number of GMPs from MLL- expression of 450% in HL60, MOLM13 and THP1 and 25–30% ENL leukemia reproduced the disease only after a significantly reduction in RS4;11 (Figure 5a). RNA interference-mediated prolonged latency period of 47 days (Po0.01, log-rank test). knockdown of MN1 expression also resulted in a significant These experiments strongly support the idea of functional decrease of clonogenic growth (in methylcellulose) of THP1 and collaboration of MN1 with MLL-ENL at the leukemic GMP cell MOLM13 and to a lesser extent of RS4;11 (Figure 5b). Similarly, population level. impaired growth potential of the cell lines expressing MN1 As MN1 has been proposed to exert an effect primarily as a siRNAs was also observed in liquid cultures (Figure 5c). The transcriptional co-regulator in hematopoietic cells, we sought to growth of HL-60 cells that express very low levels of MN1 was obtain some insights about the MN1-regulated oncogenic gene not affected by MN1 knockdown. Cell cycle analysis of THP1 expression program.2 Thus, we compared the gene expression and MOLM13 cells showed that MN1 knockdown led to a profiles of primary murine bone marrow cells overexpressing significant increase of cells in G0/G1 associated with a lower MN1 (MSCV-MN1-IRES/EYFP) with cells transduced with a number of cells in G2/M and S phase (Figure 5d). The observed control vector (MSCV-IRES/EYFP) using oligonucleotide arrays. reduced growth rate was not related to an increase of apoptotic As shown in Supplementary Table 2, transient overexpression of cell death (Supplementary Figure 3). Interestingly, siRNA- MN1 resulted in X1.5-fold upregulation of 831 genes and mediated knockdown of MN1 expression in THP1 cells was downregulation of 268 genes (Po0.05). As we asked for associated with reduced expression of the proposed MN1 target putative targets that might be involved in the leukemogenic genes CD34 and FLT3, but not with MEIS1 expression, activity of MN1, we selected genes that were more than two- suggesting that CD34 and FLT3 might also be important fold upregulated and previously functionally linked to malignant mediators of MN1 activity in human hematopoietic cells hematopoiesis and validated their expression levels using (Supplementary Figure 4). Taken together, these results showed quantitative reverse transcriptase PCR analysis. DLK1, FLT3, for the first time that increased MN1 expression provides a CD34, ANGPT2, MAP2K1IP1, HLF, PRNP and MEIS1 were positive growth signal to human leukemic cells, and functionally expressed at higher levels, and LIF and CRISP1 at lower levels in cooperates with a leukemogenic fusion oncogene through a MN1-expressing cells when compared with mock-transduced genetic program that results in expansion of the leukemia- mouse bone marrow (data not shown). We next determined initiating cells.

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 608

Figure 3 Determination of presumptive MN1 targets in hematopoietic cells. (a) Quantitative reverse transcriptase PCR (RT-PCR) analysis of the expression of selected genes in murine blasts from leukemia induced by expression of MN1, MLL-ENL or MLL-ENL þ MN1. The expression levels of target genes were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and given as DCt value relative to MOCK (MSCV-EYFP)- transduced control cells (in vitro) or bone marrow from mice transplanted with MOCK-transduced cells (in vivo). Results are mean±s.d. of triplicates from three different leukemic mice. The signiﬁcance was assessed in two ways: analysis of variance (ANOVA) tests with Bonferroni correction. (b) Quantitative RT-PCR analysis of expression of selected targets in pediatric leukemia cases with high or low levels of MN1 expression, or categorized into different subgroups of pediatric leukemia including B-ALL (n ¼ 7), infant leukemia (n ¼ 8), AML (n ¼ 6) and T-ALL (n ¼ 6). The expression levels of target genes were normalized to GAPDH, and given as DCt value relative to healthy controls (bone marrow). Dots represent mean value of duplicate for each patient sample and lines represent median levels of patient groups. The statistic analysis was performed using two-tailed t-test.

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 609 linked to carcinogenesis (Table 1).29–33 The integration near the MN1 locus resulted in significantly increased levels of mRNA expression (Figure 1a). Interestingly, the MN1 gene locus has been previously reported as a recurrent integration site in mouse models of human hematological malignancies induced by retroviral oncogene expression such as NUP98-HOXD13 or mutated AML1.20,21 Several studies suggested potential new functional links between MLL, core-binding factor and EVI1. EVI1 seems to be a downstream target in MLL-ENL and NUP98- HOX-mediated transformation.34–36 In addition, direct interactions between MLL and AML1 seem to provide epigenetic regulation of gene expression in leukemic cells.37 Although the exact molecular mechanisms remain to be elucidated, a positive role of MN1 in a putative pathway involving core-binding factor and EVI1 blocking differentiation and providing aberrant self- renewal could explain why a clone with integration-mediated elevated MN1 expression would be selected in the presence of MLL-ENL. As MN1 has in vitro and in vivo leukemogenic potential in the murine system8,9 (Figure 1), retroviral integration-mediated MN1 overexpression in MLL-ENL-induced leukemia suggested a functional cooperation. Indeed, co-expression of MN1 with MLL-ENL resulted in the induction of an aggressive AML-like phenotype after a significantly reduced latency period (Figure 1). Previous studies have proposed that in addition to hematopoietic stem cells, committed progenitor cells such as common myeloid or lymphoid progenitors (common myeloid progenitors and common lymphoid progenitors) or even GMPs can also be targets of MLL-ENL, MLL-AF9 or MOZ-TIF2 transformation.15,16,26,38 Interestingly, in normal mice the highest levels of MN1 were found in GMPs.8 Detailed comparison of the MLL- ENL-induced leukemia with the disease induced by expression of MLL-ENL and MN1 suggests that co-expression could lead to an expansion of the GMP population. Increased clonogenic activity of these cells further suggests that MN1 also regulates Figure 4 MN1 expression in human acute leukemia cell lines. (a) genes implicated in self-renewal. Accelerated induction of a MN1 mRNA expression level in human acute leukemia cell lines was leukemic phenotype in secondary transplants further supports determined using quantitative reverse transcriptase PCR (RT-PCR). the idea that GMPs represent a prominent disease-initiating Relative MN1 expression was normalized to glyceraldehyde-3- phosphate dehydrogenase (GAPDH) expression in the same sample, cell pool (also often referred as leukemic stem cells) in MLL- and standardized to HL-60 cells. Results are mean±s.d. of three ENL þ MN1 leukemia. Indeed, transplantation of sorted GMP independent experiments. (b) Detection of MN1 protein expression in cells from MLL-ENL and MLL-ENL þ MN1 leukemias into cell lines by immunofluorescence staining (Objective Â 40/1.30). secondary recipients fully supported this idea (Figure 2d). Abbreviations: MN1, MN1 recognized by MN1-specific antibody Further studies are needed to determine whether elevated bound with green fluorescence labeled secondary antibody; NORM, MN1 levels would also increase the pool of leukemia-initiating phase contrast; PI, nuclear staining by propidium iodide. cells in leukemias harboring inv16, or alterations of EVI1 that are often associated with MN1 overexpression. In adult de novo AML with no detectable cytogenetic Discussion abnormalities, MN1 expression was associated with poor response to therapy, a higher disease relapse rate and shorter Several studies have suggested that retroviral expression of MLL overall survival.10,11 These clinical observations as well as our fusion genes in the murine hematopoietic system is necessary experimental work suggest that elevated MN1 expression might but rather not sufficient to induce an acute leukemia that be a general signature for an expanded leukemic stem cell pool. recapitulates the human disease in many aspects.14–19 As our However, the nature of the disease-initiating stem cell in MLL initial screens for structural aberrations (not shown) did not fusion-induced murine leukemia is still a matter of debate: the revealed any aberrations, we focused on proviral integration expression of Gr1/Mac1 þ and c-Kit þ covering 25% of the sites as potential cooperating events in 10 mice with MLL-ENL- blasts as well as lineage-negative (o1% of leukemic blasts) induced acute leukemia. We were able to clone and validate leukemic-hematopoietic stem cells, L-GMPs or L-common 25 integrations in two independent experiments. It has been lymphoid progenitors (CLPs) that have been proposed to exert previously shown that mouse leukemia genes identified by an effect as leukemic stem cells.15–17,26 However, independent retroviral insertion mutagenesis are more frequently differen- of the chosen phenotypic markers, we observed a significant tially expressed in human acute leukemia than randomly increase in potential leukemic stem cell pool when MN1 was selected genes or genes located more distantly from a provirus co-expressed with MLL-ENL, as illustrated by secondary integration.27,28 Interestingly, several genes flanking the transplantation experiments (Figure 2). Interestingly, transient integration sites, such as SOCS1, JARID2, PTPRE, HSP90, and stable expression of MN1 was associated with increased GADD45, MGAT5, PP2R5C or MN1, have been previously CD34 expression (Figure 3). Similarly, gene expression profiling

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 610

Figure 5 Knockdown of MN1 expression impairs the growth of human acute leukemia cells with elevated MN1 levels. (a) Efficacy of the lentiviral-mediated MN1 knockdown by human MN1-specific siRNA (black bars) or scramble siRNA (white bars) assessed using quantitative reverse transcriptase PCR (RT-PCR). Results are mean values±s.d. of three independent experiments. The significance was determined using two- tailed t-test (*Po0.05, **Po0.01). (b) MN1 knockdown impaired clonogenic growth (in methylcellulose) of THP1, MOLM13 and RS4;11 cell lines expressing high levels of MN1 but not HL-60 expressing very low levels of MN1. Results are mean±s.d. of three independent experiments. The significance was determined using two-tailed t-test (*Po0.05, **Po0.01). (c) MN1 knockdown impaired proliferation of THP1, MOLM13 and RS4;11 cell lines but not HL-60 in liquid culture. Results are mean±s.d. of three independent experiments. The significance was determined using two-tailed t-test (*Po0.05, **Po0.01). (d) Cell cycle analysis revealed that MN1 knockdown resulted in increased cells in G0/G1 phase, and less cells in S and G2/M phase.

of cytogenetically normal adult AML cases also very recently currently being validated in a larger cohort of pediatric acute revealed an association between MN1 and CD34 expression leukemia patients. levels.11 CD34 is a well-characterized marker of early hemato- Our results suggest a functional cooperation of MN1 with MLL- poietic progenitor cells.39 We revealed, for the first time, that a ENL leukemogenesis. It is worth noting that not all studied MLL significant correlation of MN1 expression with CD34 levels was fusion-positive human leukemia cell lines expressed elevated not only observed in the mouse leukemia model but also in a MN1 levels (Figure 4). In addition, our own and other researchers’ small cohort of primary samples from pediatric patients with gene expression profiling studies of experimental models of MLL B-cell ALL or infant leukemia (Figure 3). These observations are fusion-induced leukemia never revealed MN1 as being a primary

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 611 upregulated target17,26 (T Liu, D Jankovic, J Schwaller et al, myeloid cells and causes myeloid malignancy in mice. Leukemia unpublished observation). This raised the possibility that the MN1 2006; 20: 1582–1592. expression status in a leukemic blast might be dependent on the 7 Valk PJ, Verhaak RG, Beijen MA, Erpelinck CA, Barjesteh van maturation grade of the cell that was initially targeted by the MLL Waalwijk van Doorn-Khosrovani S, Boer JM et al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J fusion and/or on a currently unknown mechanism that would Med 2004; 350: 1617–1628. maintain the MN1 levels high. Interestingly, an unusual methyla- 8 Carella C, Bonten J, Sirma S, Kranenburg TA, Terranova S, Klein- tion pattern of the MN1 promoter region was recently reported, Geltink R et al. MN1 overexpression is an important step in the suggesting that epigenetic mechanisms might be needed to sustain development of inv(16) AML. Leukemia 2007; 21: 1679–1690. the elevated MN1 expression levels in cell transformed by 9 Heuser M, Argiropoulos B, Kuchenbauer F, Yung E, Piper J, Fung S oncogenic MLL fusion genes.40 et al. MN1 overexpression induces acute myeloid leukemia in mice and predicts ATRA resistance in patients with AML. Blood Forced siRNA-mediated downregulation of MN1 expression 2007; 110: 1639–1647. in cells that overexpress MN1 impaired growth in methylcellu- 10 Heuser M, Beutel G, Krauter J, Dohner K, von Neuhoff N, lose (colony formation) as well as in liquid culture. MN1 Schlegelberger B et al. High meningioma 1 (MN1) expression as a knockdown was associated with cell cycle arrest in G0/G1 phase predictor for poor outcome in acute myeloid leukemia with normal (Figure 5). These results show, for the first time, that aberrant cytogenetics. Blood 2006; 108: 3898–3905. MN1 expression could be an important factor for proliferation of 11 Langer C, Marcucci G, Holland KB, Radmacher MD, Maharry K, Paschka P et al. Prognostic importance of MN1 transcript levels, leukemic cells. As leukemic and normal hematopoietic cells and biologic insights from MN1-associated gene and microRNA expressing normal levels of MN1 were not affected by siRNA- expression signatures in cytogenetically normal acute myeloid mediated knockdown, selective inhibition of MN1 in leukemia- leukemia: a Cancer and Leukemia Group B Study. J Clin Oncol initiating cells might be possible without affecting normal tissue. 2009; 27: 3198–3204. Currently, ongoing structure–function studies as well as protein 12 Gilliland DG, Jordan CT, Felix CA. The molecular basis of interaction screens to search for critical domains for the leukemia. Hematology Am Soc Hematol Educ Program 2004, leukemogenic activity of MN1 will provide new insights for 80–97. 13 Meyer C, Schneider B, Jakob S, Strehl S, Attarbaschi A, Schnittger S future therapeutic targeting of the aberrant MN1 activity. et al. The MLL recombinome of acute leukemias. Leukemia 2006; 20: 777–784. 14 Krivtsov AV, Armstrong SA. MLL translocations, histone modifica- Conflict of interest tions and leukaemia stem-cell development. Nat Rev Cancer 2007; 7: 823–833. The authors declare no conflict of interest. 15 Cozzio A, Passegue E, Ayton PM, Karsunky H, Cleary ML, Weissman IL. Similar MLL-associated leukemias arising from self- renewing stem cells and short-lived myeloid progenitors. Genes Acknowledgements Dev 2003; 17: 3029–3035. 16 Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J et al. We thank Verena Ja¨ggin, Emmanuel Traunecker and Ulrich Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature 2006; 442: 818–822. Schneider for technical help, and Radek Skoda for carefully 17 Somervaille TC, Cleary ML. Identification and characterization of reading the manuscript. This work was supported by the leukemia stem cells in murine MLL-AF9 acute myeloid leukemia. Gertrude von Meissner Foundation and the Swiss National Cancer Cell 2006; 10: 257–268. Science Foundation (3100AO-116587) to JS, and the ‘Associa- 18 Wong P, Iwasaki M, Somervaille TC, So CW, Cleary ML. Meis1 is zione Italiana Ricerca sul Cancro’AIRC and by ‘Fondazione an essential and rate-limiting regulator of MLL leukemia stem cell Cariplo’ to AB. potential. Genes Dev 2007; 21: 2762–2774. 19 Horton SJ, Walf-Vorderwulbecke V, Chatters SJ, Sebire NJ, de Boer J, Williams O. Acute myeloid leukemia induced by MLL-ENL is Contributions cured by oncogene ablation despite acquisition of complex TL, DJ, LB, SE, FB, VS and VR performed research; AB analyzed genetic abnormalities. Blood 2009; 113: 4922–4929. data; and JS performed research, analyzed data and wrote the 20 Slape C, Hartung H, Lin YW, Bies J, Wolff L, Aplan PD. Retroviral paper. insertional mutagenesis identifies genes that collaborate with NUP98-HOXD13 during leukemic transformation. Cancer Res 2007; 67: 5148–5155. 21 Watanabe-Okochi N, Kitaura J, Ono R, Harada H, Harada Y, References Komeno Y et al. AML1 mutations induced MDS and MDS/AML in a mouse BMT model. Blood 2008; 111: 4297–4308. 1 Lekanne Deprez RH, Riegman PH, Groen NA, Warringa UL, van 22 Schwaller J, Frantsve J, Aster J, Williams IR, Tomasson MH, Ross TS Biezen NA, Molijn AC et al. Cloning and characterization of MN1, et al. Transformation of hematopoietic cell lines to growth-factor a gene from chromosome 22q11, which is disrupted by a balanced independence and induction of a fatal myelo- and lymphoproli- translocation in a meningioma. Oncogene 1995; 10: 1521–1528. ferative disease in mice by retrovirally transduced TEL/JAK2 fusion 2 Grosveld GC. MN1, a novel player in human AML. Blood Cells genes. EMBOJ 1998; 17: 5321–5333. Mol Dis 2007; 39: 336–339. 23 Lavau C, Szilvassy SJ, Slany R, Cleary ML. Immortalization and 3 Buijs A, Sherr S, van Baal S, van Bezouw S, van der Plas D, Geurts leukemic transformation of a myelomonocytic precursor by van Kessel A et al. Translocation (12;22) (p13;q11) in myeloproli- retrovirally transduced HRX-ENL. EMBOJ 1997; 16: 4226–4237. ferative disorders results in fusion of the ETS-like TEL gene on 12p13 24 Du Y, Jenkins NA, Copeland NG. Insertional mutagenesis to the MN1 gene on 22q11. Oncogene 1995; 10: 1511–1519. identifies genes that promote the immortalization of primary bone 4 Kawagoe H, Grosveld GC. Conditional MN1-TEL knock-in mice marrow progenitor cells. Blood 2005; 106: 3932–3939. develop acute myeloid leukemia in conjunction with overexpres- 25 Du Y, Spence SE, Jenkins NA, Copeland NG. Cooperating cancer- sion of HOXA9. Blood 2005; 106: 4269–4277. gene identification through oncogenic-retrovirus-induced inser- 5 Kawagoe H, Grosveld GC. MN1-TEL myeloid oncoprotein tional mutagenesis. Blood 2005; 106: 2498–2505. expressed in multipotent progenitors perturbs both myeloid and 26 Chen W, Kumar AR, Hudson WA, Li Q, Wu B, Staggs RA et al. lymphoid growth and causes T-lymphoid tumors in mice. Blood Malignant transformation initiated by Mll-AF9: gene dosage and 2005; 106: 4278–4286. critical target cells. Cancer Cell 2008; 13: 432–440. 6 Carella C, Bonten J, Rehg J, Grosveld GC. MN1-TEL, the product of 27 Erkeland SJ, Verhaak RG, Valk PJ, Delwel R, Lowenberg B, Touw the t(12;22) in human myeloid leukemia, immortalizes murine IP. Significance of murine retroviral mutagenesis for identification

Leukemia Meningioma 1 as collaborating leukemogenic oncogene T Liu et al 612 of disease genes in human acute myeloid leukemia. Cancer Res 34 Palmqvist L, Pineault N, Wasslavik C, Humphries RK. Candidate 2006; 66: 622–626. genes for expansion and transformation of hematopoietic stem 28 Touw IP, Erkeland SJ. Retroviral insertion mutagenesis in mice as a cells by NUP98-HOX fusion genes. PLoS ONE 2007; 2: e768. comparative oncogenomics tool to identify disease genes in 35 Barjesteh van Waalwijk van Doorn-Khosrovani S, Erpelinck C, van human leukemia. Mol Ther 2007; 15: 13–19. Putten WL, Valk PJ, van der Poel-van de Luytgaarde S, Hack R 29 Watanabe D, Ezoe S, Fujimoto M, Kimura A, Saito Y, Nagai H et al. High EVI1 expression predicts poor survival in acute myeloid leukemia: a study of 319 de novo AML patients. Blood 2003; 101: et al. Suppressor of cytokine signalling-1 gene silencing in acute 837–845. myeloid leukaemia and human haematopoietic cell lines. Br J 36 Arai S, Goyama S, Shimabe M, Ichikawa M, Imai Y, Takahashi T Haematol 2004; 126: 726–735. et al. Evi-1 is a direct target of MLL oncoproteins in hematopoietic 30 Lau KS, Dennis JW. N-Glycans in cancer progression. Glycobiology stem cells. Blood (ASH Annual Meeting Abstracts) 2008; 112: 282. 2008; 18: 750–760. 37 Huang G, Elf S, Yan X, Wang L, Liu Y, Sashida G et al. Previously 31 Tanuma N, Nakamura K, Shima H, Kikuchi K. Protein-tyrosine unknown interactions between AML1 MLL provide epigenetic phosphatase PTPepsilon C inhibits Jak-STAT signaling and regulation of gene expression in normal hematopoiesis in differentiation induced by interleukin-6 and leukemia inhibi- leukemia. Blood (ASH Annual Meeting Abstracts) 2008; 112: 282. tory factor in M1 leukemia cells. J Biol Chem 2000; 275: 38 Huntly BJ, Shigematsu H, Deguchi K, Lee BH, Mizuno S, Duclos N 28216–28221. et al. MOZ-TIF2, but not BCR-ABL, confers properties of leukemic 32 Perrotti D, Neviani P. ReSETting PP2A tumour suppressor activity stem cells to committed murine hematopoietic progenitors. Cancer Cell 2004; 6: 587–596. in blast crisis and imatinib-resistant chronic myelogenous leukaemia. 39 Krause DS, Fackler MJ, Civin CI, May WS. CD34: structure, Br J Cancer 2006; 95: 775–781. biology, and clinical utility. Blood 1996; 87: 1–13. 33 Flandrin P, Guyotat D, Duval A, Cornillon J, Tavernier E, Nadal N 40 Kuang SQ, Tong WG, Yang H, Lin W, Lee MK, Fang ZH et al. et al. Signiﬁcance of heat-shock protein (HSP) 90 expression in Genome-wide identiﬁcation of aberrantly methylated promoter acute myeloid leukemia cells. Cell Stress Chaperones 2008; 13: associated CpG islands in acute lymphocytic leukemia. Leukemia 357–364. 2008; 22: 1529–1538.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia